Factories are too DULL for Google's robo-dreams: Behold the GATAMAMs
'Look on my works, ye Mighty, and Search On!'
Comment Google's spree of robot acquisitions represents a major bet by the ad giant that it can wring more data from the world, and make some money that's not from advertising along the way. And, we reckon, it will deal with a serious problem lurking in the company's future.
When news broke on Friday that Google had bought DARPA's favorite robot company Boston Dynamics, the reactions were a mixture of confusion and obligatory Skynet-assertions – Google! Robots! What?
But when you think about the internet giant's goal of "organizing the world's information", the acquisition makes total sense.
Google isn't so much a company as an empire that is fueled by data. For this reason, the web search king has a natural inclination to have computers perform as many tasks as possible because they can be far more reliable than squishy humans. After all, computers are better recorders of data than people, trained neural networks are more effective categorizers, and robots have a chance of being more dependable collectors.
Though it's claimed Google may be spinning up a robot wing to sell packaging machines to warehouses, and to improve its own distribution and manufacturing capabilities, we suspect that its real long-term goal is to develop new ways of sucking up more data from the real world more cheaply.
Already, the company's main products – Google Search, YouTube, and the Android operating system – are little more than an interface for translating desires (Search), dreams (YouTube), and actions (Android) of flesh-and-blood creatures into structured, analyzable digital information.
Now that Google has effectively corralled us into its digital domain – and understood us well enough to serve targeted videos, adverts and other stuff – it is faced with the conundrum of finding new markets to dominate.
Advertising is a lucrative market, but given Google's capabilities and vast profits ($2.97bn in its last quarter, for instance), its engineers will be eager to use that might to plunder new seams. One such moonshot area is self-driving cars, which are backed by heavy investment into Google Maps and various sensor technologies.
But amid its robot gobbling spree skulks a strange specter that threatens to linger long into the future. First, let's look at Google's acquisitions to date in the robotics world:
- Makani Power (airborne wind turbines)
- Schaft, Inc (super-strong bipedal robots)
- Bot and Dolly (automated camera systems)
- Industrial Perception Inc (computer vision systems and arms for loading and unloading equipment from trucks)
- Redwood Robotics (robot butlers – yes, really)
- Meka (MIT-spinoff chiefly known for its "human-scale robot technologies")
- Holomni (omnidirectional wheels)
- Boston Dynamics (agile robots capable of traveling through rugged areas)
Individually, these are all interesting purchases, and we can foresee a number of neat innovations through the combination of a few of them, including quality inspection at point of automated deliver (Industrial Perception, Inc and Shaft, Inc), personal nursing systems (Redwood Robotics and Meka and Holomni), hazard response drones (Shaft and Bot and Dolly and Makani), and many others.
But they also hint at a far-off creation that we will call the Google All Terrain Automated Mapper and Analyzer and Manipulator (GATAMAM).
The GATAMAM would have the rugged all-terrain capabilities of Boston Dynamics, combined with the dexterity of Redwood Robotics and strength of Shaft Inc, and the intensely clever software systems of Industrial Perception and Meka. It would be coupled with a Makani Power wind turbine outfitted with sensor technology derived from Bot and Dolly and Industrial Perception Inc.
This GATAMAM would directly deal with a problem that lurks in Google's future – info-rot.
Just as Android lets Google freely harvest vast quantities of data about the day-to-day habits of people, and Google Search lets it turn the intent of its hundreds of millions of users into advertising dollars, an upcoming army of Google Robots will let the company draw more data – photos, signs, video, sounds, temperatures, smells, you name it – from the physical world, and at a lower cost.
Though Google harvests many petabytes of data a year from meatspace in the form of its Street View mapping arsenal, there are many unmapped and unplumbed parts of the world. For Google to expand its services such as maps, self-driving cars, and automated shopping into new markets, it will need data.
This data costs a lot to gather and requires the company to employ an army of humans to take Street View boats, backpacks, cars, snowmobiles, bicycles and other forms of camera-fitted transportation throughout the world. By deploying an army of GATAMAMs Google could cut the cost of gathering data, and be able to update its store of data more regularly to keep up with changes in the physical world – even in the skies above us, the water surrounding us, and in the ground below.
The more data Google can ingest, and the more frequently it can update it, the more beguiling it can make its free services (Search, Gmail, Maps, Android, YouTube, and so on), thus securing our ongoing personal data donations that it converts into revenue.
By having an army of technologically advanced GATAMAMs pulling in data from the world, Google would be able to confront one of the greatest threats to its business, and give itself a planet-spanning legion of mobile, semi-automated sensors to suck in information to train its vast artificial intelligence systems on.
It will also help Google deal with "The Great Wall of China" paradox which currently threatens its business.
This paradox is that for China to protect itself from the Mongols, it was motivated to build a vast wall that kept the enemy out, and could be patrolled by an army. The bigger the wall got, the more resources had to be devoted to its maintenance, and the more politically unfeasible it became to propose any new ideas for how to keep the Mongols out due to the vested careers and political livelihoods of many, many people.
In Google's world, the Mongols are the lurking threat of outdated data creeping in and destroying the value of its servers, and its Wall of China is its ability to effectively take in data from the fringes of its empire and use this to sustain its core without bankrupting itself.
Without an automated army of info-slurping robots, Google's costs will climb and it will become wedded to the thousands upon thousands of people needed to sustain its maps and wider library of information, and their costs and limits and (hiss!) unions.
With a mechanized army, its top engineers and executives would only be beholden to themselves. It's enough to make you wonder. ®
Sponsored: Benefits from the lessons learned in HPC